EXR Confusion not fully resolved

Started 7 months ago | Discussions thread
Timur Born
Senior MemberPosts: 3,843
Like?
Re: EXR Confusion not fully resolved
In reply to photoreddi, 7 months ago

photoreddi wrote:

I've seen that control before but not here. Do you see it here?

It's not available for the ISO measurement. Look at SN 18% and DR measurements and you will see the control.

Wow. You're still not acknowledging the use of different exposures by the two sensor halves in EXR DR photos.

That's because I took it as a given basis for our discussion. Yes, in EXR DR half of the sensor is exposed 1 (DR200) or 2 (DR400) stops shorter (in time) than the other half. Read my age old user article "Summary of Fujifilm X10's functions and issues" to get an idea of my understanding (partly out-dated, but the basics of what you wanted me to write out are all in there).

And in practice taking two 6 mp images with different shutter speeds at the same time is nearly the same as taking two consecutive 6 mp images merged in post. The former has the advantage with alignment of handheld shooting and is slightly less susceptible to in-frame motion. On top of that the Bayer pattern of averaged photosites *before* demosaicing and downsizing is claimed to be superior to downsizing *after* demosaicing (non EXR cams). The latter sounds reasonable, but isn't so easy to test with the X10 due to its unique sensor size compared to other 12 mp cameras out there.

But taking two 6 mp images at the same time offers less resolution and produces more noise than taking two 12 mp consecutive images merged in post. The problem with taking two half-resolution images of very high dynamic range scenes is that half the sensor is clipped white in highlights and half the sensor is clipped black in shadows. This effectively reduces SN ratio of both these areas to the equivalent of taking an image with only half the sensor size. This is because clipped information is non-information other than bright (max) and dark (0). So EXR DR is useful, but comes with its own limitations that many people here don't seem/want to recognize.

I posted this example several times in the past. As can be seen EXR DR 400 at ISO 100 is more noisy in the highlights than HR DR 400 at ISO 400. That's not even a surprise, because the HR DR 400 shot internally uses ISO 100 equivalent gain, exposes for the highlights (2 stops less exposure time due to nominal ISO 400) and then only pushes shadows and mids digitally to ISO 400 equivalent noise. The EXR DR 400 / ISO 100 shot on the other hand suffers from half the photosites being clipped pure white and thus not providing any useful information other than "bright".

Here is a more emphasized version to make the difference more visible.

I wrote this in another thread: If "highlight protection" is what you care for then properly expose for the highlights. Most people on here seem to equate "highlight protection" with "dynamic range", but they are not the same. EXR DR protect from user errors, but for scenes where you (r images) don't need the extended dynamic range you get worse image quality without benefits.

Do you really not understand this after so much time has passed? I guess that you and Trevor both subscribe to the "party trick" theory after all.

You seem to associate my arguments too much with what you perceive as Trevor's view on things. Since I don't want to repeat all what I have written in the past just to give you a better idea of my understanding of EXR it may be best to read that old article and search for some old posts or just ask me specific questions where confusion arises. At least that's better than putting statements in my mouth that I did not make.

-- hide signature --

Red flash eyes save lives and eye-sight!
http://en.wikipedia.org/wiki/Retinoblastoma

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow